Enhancing the mechanical and anticorrosion properties of 316L stainless steel via a cathodic plasma electrolytic nitriding treatment with added PEG

1. Introduction

Although the corrosion resistance of 316L stainless steels is acceptable, the wear resistance and hardness are limited for industrial applications. To improve these mechanical properties without destroying the corrosion resistance of austenitic stainless steels, surface hardening processes, such as carburizing and nitriding, were used [[1], [2], [3], [4], [5], [6]]. A number of studies reported that conventional gas nitriding and ion nitriding technologies were able to form a 2-5 μm compound layer and a 2-20 μm diffusion layer on the surface of the sample, and the obtained nitrided layers exhibited good mechanical properties and corrosion resistance [[1], [2], [3]]. For gas nitriding technologies, due to the limitations of reaction kinetics and thermodynamics, the process duration is generally 40-90 h in the temperature range of 460-530 °C, which is considered to be relatively long. This condition may cause deformation and failure of the workpieces during the nitriding process [2]. The surface hardness of the workpiece after ion nitriding or heat treatment may not meet the technical requirements, leading to a rework or even scrap [3,4]. Moreover, the aforementioned nitriding technologies need to be carried out in a high-vacuum environment and last a relatively long time.

In recent decades, cathodic plasma electrolytic nitriding (CPEN) has shown broad application prospects, and this process is conducted in an atmospheric environment and quenched directly in the electrolyte within a few minutes. The acquired surface hardness is promising as well as wear and corrosion resistant [[7], [8], [9], [10], [11]]. The plasma film on the metal/electrolyte interface is produced on the cathode that is in the electrolyte and generates thermochemical energy and electric field energy on the surface of the sample [30]. The CPEN techniques, which are more efficient than traditional nitriding processes, have a 3-6 μm/min growth rate and form a N diffusion layer on carbon steel; the CPEN technique then produces a 15-60 μm thickness layer after 5-10 min consisting of expanded austenite (γ_N) [[12], [13], [14], [15]]. Studies [14,15] have shown that urea solutions are applicable to nitriding because of the formation of active nitrogen from the ammonia, which is produced by the cathodic reactions. However, because of the random discharge phenomenon at the metal/electrolyte interface, the compound layer obtained in conductive urea salt (e.g., sodium carbonate or ammonium chloride) solutions is too thin (approximately 1-2 μm) [26,29]. Therefore, the composition of the solution should be changed to promote a uniform discharge on the surfaces of the samples in solution and improve the penetration efficiency. According to the disruptive discharge principle [31], the liberation of the gas envelope and discharge of the plasma are both affected by the bath voltage and thickness of the gas envelope. Their molecules of some nitriding agents (e.g., urea and formamide) that participate in the formation of a plasma are not large enough to bind to the gas at the interface of the metal/electrolyte and make it difficult to control the thickness of the gas envelope for the production of a large uniform volume discharge.

PEG is a kind of macromolecular soluble substances that can increase the viscosity of a solution, and some studies have reported that PEG was conducive to controlling the thickness of the gas envelope on the surface [24,26]. In this investigation, PEG2000 was added to the solution to bind the gases between the plasma film and the surface of steel and to focus the nitriding energy so that a thick nitrided layer could be formed. The mechanical and corrosion properties of the obtained layer were then characterized.

2. Experimental

For the CPEN process, 316L stainless steel specimens made by Baosteel (10 mm × 10 mm × 2 mm) with a chemical composition (wt%) of 0.015 C, 0.5 Si, 0.8 Mn, 0.012 P, 0.07 S, 2.3 Mo, 12.6 Ni, 17.15 Cr and balance Fe were attached to the negative terminal of a DC power supply (maximum voltage: 500 V) and used as the cathode. A platinum plate electrode (110 mm × 40 mm × 2 mm) was used as the anode. Previously, all specimens were polished with 800 grit silicon carbide abrasive paper, thoroughly degreased by ethanol, washed in deionized water and then dried in lab air.

In the electrolytic bath, the applied voltage was increased from 0 to 190 V at a rate of 1 V s^-1, with a duration of 10 min. The current density-voltage (i_c-V) curves in the same concentration of urea aqueous solution (Solution C listed in Table 1) were recorded by the computer. There was a comparison blank test and two tests for focusing the nitriding energy on the surface of the steel: I: the sample was nitried with six faces; II: five surfaces of the sample were sealed using AB glue and silica gel and only one surface was exposed; and III: based on experiment I, the electrolyte was added with 10 g l^-1 PEG2000. The experiments were carried out at 10 °C, which was controlled by circulating cooling water.

To investigate the effects of urea concentration in the process of CPEN, all specimens were treated in the urea aqueous solutions shown in Table 1 that had different concentrations and were denoted as A, B, C, or D.

After treatment in the electrolytes with different concentrations of urea, the specimens were sealed with an epoxy resin and etched in a 5% ferric chloride-hydrochloric acid solution. The cross-sectional morphology and composition of the nitrided layers were investigated by optical microscopy (OM) and scanning electron microscopy (SEM, Quanta 250), where an attached energy dispersive spectroscope was used to conduct energy dispersive spectroscopy (EDS). The thickness of each nitrided specimens was measured thrice by built-in software of SEM, and the standard deviation were given after the mean value. The standard deviation is calculated as follows:

σ=$\sqrt{\frac{1}{N}∑\limits_{i=1}^{n}(x_i-μ)^2}$ (1)

where σ is the standard deviation, xi is the thickness measured each time, N stands for times of measurement that equals to 3, μ stands for the mean value of three measurements.

The X-ray diffraction (XRD, Smart Lab) was performed to determine the phase compositions of the nitrided layer with Cu Kα radiation in the incident angle range of 10°-100° at 40 kV and 200 mA with a step size of 0.02°.

Dry sliding wear tests were conducted on an untreated specimen and a nitrided specimen by a tribometer (UMT2, CETR) to gain the coefficient of friction (COF) of each nitrided specimen. A 3 N load was applied for 15 min within a distance of 15 mm by an Al₂O₃ ball in an air atmosphere at room temperature, and the friction force was monitored to obtain the coefficient of friction. The worn surfaces were analyzed by three-dimensional confocal microscopy (VHX-2000) to get the wear rate. The microhardness of the cross-sectional nitrided specimens was measured by a Vickers microhardness tester (HXD-1000) under a load of 100 g for 15 s.

The nitrided layers were evaluated by potentiodynamic polarization tests after being washed in deionized water and then dried in lab air. Electrochemical polarization curves of four nitrided specimens and a bare specimen were tested with a working area of 1 cm² using a PGSTAT302N electrochemical workstation (Metrohm, Switzerland) in 3.5% NaCl solution at room temperature. The three-electrode electrochemical device consisted of the specimen as the working electrode, a saturated calomel electrode (SCE) as the reference electrode and a platinum sheet as the auxiliary electrode. The scan rate was set at 0.5 mV s^-1, and the scanning potential was in the range of -0.4 to 0.8 V_ocp (OCP: open current potential).

3. Results and discussion

3.1. Discharge mechanism of CPEN after adding PEG

Fig. 1 shows that electrolytic nitriding is a complex physical and chemical process. The following chemical reactions occurred on the surface of the sample [5,15,[19], [20], [21]]:

H₂O↔H⁺+OH^- (2)

Fe_solid+NH_x,gas→FeNH_2-3+Fe_2-4N+N₂↑+H₂↑+NH₃↑ (3)

NH₄Cl+H₂O↔NH₄OH+HCl (4)

(NH₂)₂CO+H₂O→NH₃↑+HNCO+H₂↑+CO₂ (5)

View original graphic| Download| PPT slide

Fig. 1.   Schematic diagram of the cathodic plasma electrolytic nitriding process.

Previous studies [15,24] showed that among the reaction products, carbon dioxide molecules were cohesive and difficult to break down, so the gases for the plasma formation were mainly H₂ and NH₃. With increasing voltage, the gas film became sufficiently stable and then a uniform discharge occurred, improving the electric field energy. The poor conductivity gases caused an electrical breakdown at the spark voltage and formed iron compounds (e.g., ε-Fe_2-3N, γ’-Fe₄N, and FeN_0.076) and a diffusion zone [[16], [17], [18], [19], [20],22,23,[37], [38], [39], [40]]. Generally, the gas film was too thin and sparse to produce a uniform plasma film, and the nitriding energy was not large enough to increase the nitrided layer thickness. The spark voltage (U_spark) is a critical parameter for revealing the nitriding ability, and the gas film thickness has significantly influence on the dynamics of nitrogen diffusion. The discharge processes of cathodic plasma electrolytic were investigated via i_c-V curves to describe the kinetic properties [25,27]. The essential criterion for plasma discharge can be expressed as E ≥ E_{spark ignition,} where the E_{spark ignition} stands for critical breakdown electric field intensity of gases gathering on the surface of sample. The value of E can be calculated by E = V/δ, where δ and V represent the thickness of gas film and the bath voltage, respectively. It can be inferred that the plasma discharge behaviour has a significant effect on the film thickness, and if the value of δ increases, the value of E increases as the voltage increases. Fig. 2 shows the behaviour of the PEG binding gas film generated on the sample surface. From the PEG structural formula, it can be seen that the molecule does not contain nitrogen atoms. Therefore, PEG2000 is not an osmotic agent and is suitable as an additive agent to control the i_c and for investigating the effects of urea concentration in the process of CPEN.

View original graphic| Download| PPT slide

Fig. 2.   Schematic diagram of PEG to bond film on the metal/electrolyte interface.

Fig. 3 shows i_c-V curves for samples nitrided in the solution with a urea concentration of 543 g l^-1. It can be seen that the spark ignition current density (i_spark) decreased greatly for experiments II and III mentioned above, and the decrease in experiment III was more significant. A decrease in the value of i_spark represents an increase in the plasma film charge resistance, which can indicate that the plasma film increased in thickness and density. The value of U_spark increased with a decrease in i_spark, as shown in Fig. 3. The temperature is a key factor in the nitriding process, and the value of U_spark reflects the nitriding energy in a liquid. The improvement of U_spark can increase the nitriding temperature and the electric field intensity. Thus, a high U_spark can provide sufficient energy for the infiltration of nitrogen atoms from the kinetics to realize a rapid nitriding process. Therefore, the addition of PEG2000 can enhance the ability of the electrolyte to bind the gas to the film with increasing the value of δ, in the meantime, increase the electric field strength to improve the nitriding energy.

View original graphic| Download| PPT slide

Fig. 3.   Current density-voltage (i_c-V) curves of three experiments for investigating the increase of nitriding energy in CPEN.

3.2. Morphology and composition of the nitrided layers

Fig. 4 shows the optical micrographs of cross-sections of the samples nitrided by the three experiments mentioned above that focused the nitriding energy. The nitrided layers that were resistant to etching by the ferric chloride-hydrochloric acid solution appeared bright under an optical microscope. Meanwhile, the thickest nitrided layer was obtained by method III compared to methods I and II. Because the thickness of the nitrided layers is positively associated with the plasma charge energy, the escape of the gas film from the surface was impeded by adding PEG2000 to the electrolyte, which enhanced the nitriding energy and presented a decrease of the current density in i_c-V curve.

View original graphic| Download| PPT slide

Fig. 4.   Cross-sectional morphologies obtained from three experiments for focusing the nitriding energy: (a) I: sample without any pretreatments; (b) II: five surfaces of the sample were sealed; (c) III: PEG2000 was added to the solution.

The cross-section images of the nitrided samples that underwent experiments with different electrolytes were analysed by SEM and EDS, and the results are shown in Fig. 5. It was difficult to identify the interface between the compound zone and the diffusion zone in the cross-section images, and the nitrogen concentration of the nitrided layer treated in solution C was much higher than that of the matrix. The SEM images show that the thickness of the eroded nitrided layers treated in solutions A, B, C, and D can be distinguished from each other. The change in the mean thickness is positively related to the urea concentration, as shown in Table 2.

View original graphic| Download| PPT slide

Fig. 5.   SEM cross-section images (a-d) of samples treated in solutions A, B, C and D after erosion and (e) SEM and EDS analysis of the sample treated by method III without erosion.

The XRD patterns of the nitrided samples and steel substrate are shown in Fig. 6. It can be seen that the austenite peaks of the nitrided samples treated in solution A and B shifted slightly (∼2°) towards low diffraction angles, which was caused by the expanded austenite (γ_N), a kind of solid solution. There was a certain amount of iron nitride (FeN_0.076) on the surface of the nitrided layer treated in solution C. The FeN_0.076 phase with a face-centred cubic (fcc) structure was a relatively stable phase with great ductility (B/G ≈ 3 > 1.75; B: bulk modulus; G: shear modulus) due to the strong covalent bonding between the Fe and N atoms [28]. Thus, the peak intensities of FeN_0.076 increased as the concentration of urea increased. It was verified that FeN_0.076 existing as a nanophase promotes the elasticity, hardness and corrosion resistance of the nitrided layer, and these conclusions were confirmed in subsequent tests. In contrast to other solutions, when the concentration of urea reached 543 g l^-1 (i.e., solution D), Fe₃O₄ peaks were observed instead of γ_N and iron nitride peaks, and the Fe₃O₄ peak intensities were obviously stronger than those of the austenite phase. It can be inferred that the formation of Fe₃O₄ is related to the high temperature oxidation by water vapour combined with the initial charge process, in which the gas film is mainly composed of steam produced by the ohmic heating of water. It is noted that the nitrides were undetected in the layers formed in decreased urea concentration in solutions A and B; thus, one can speculate the concentration of urea has a certain influence on the nitriding temperatures and affects the nitriding reaction-diffusion kinetics.

View original graphic| Download| PPT slide

Fig. 6.   XRD patterns of the untreated samples and samples treated in solutions A, B, C and D.

The formation of substances in the nitrided layer can be explained according to the Ginstling equation [30] and the Arrhenius equation:

$\frac{dx}{dt}=\frac{D}{ε}(\frac{∂c}{∂r})$ (6)

K=Ae$\frac{-E_a}{RT}$ (7)

where dx/dt is the formation rate of the layer (μm/s), x is the layer thickness, and t is the time; D and ε are the diffusivity (μm² s^-1) of N atoms in the plasma and a proportional constant, respectively; and ∂c/∂r is the concentration gradient (mo μm^-1) of N atoms. The concentration of N atoms in the layer is significantly influenced by the value of ∂c/∂r and D, which are both affected by the reaction temperature (T). In Eq. (7), K (chemical reaction rate constant, min^-1) is exponentially related to 1/T. The values of D and K increase as T increases at the interface between the sample and solution. Thus, the inconspicuous compound layers obtained in solutions A and B may be caused by the decreased diffusion rate of nitrogen at low temperatures. In this way, the plasma charged energy may not be sufficient for ion implantation and thus it is difficult to precipitate Cr_xN_y or Fe₄N nitrides and even low-nitrogen nitrides (FeN_0.076) on the surface.

The above findings further indicate that urea molecules could hinder the diffusion of oxygen and urea molecules so that those molecules are encapsulated at the metal/electrolyte interface in an electrolyte with a high urea concentration. When PEG molecules are on the surface, the temperature of the gas film increases quickly. Therefore, the surface of the sample could be oxidized at the elevated temperature, resulting in an increase in the oxygen on the surface of the nitrided layer and the generation of Fe₃O₄. When also considering Fig. 2, it can now be confirmed that the addition of PEG2000 promoted the gas film to be bonded to the cathode, which reduced the current density and increased the volume density of the elements in the plasma. Therefore, the nitride layer thickness can be increased due to the increase in the urea concentration and addition of PEG.

According to the XRD patterns in Fig. 6, the width of the γ_N diffraction peaks increased and shifted towards a lower diffraction angle, which was also observed during low-temperature nitriding or carburizing of stainless steel [[5], [6], [7], [8]]. During nitriding, the active particles infiltrated into the austenite face-centred cubic lattice of the stainless steel and led to expansion of the lattice. In addition, the shift of the diffraction peak to a lower angle is also related to the residual stress of the film and presence of defects, such as stacking faults in the nitride layer, which do not substantially improve the hardness value.

3.3. Mechanical properties analyses

Among all solutions considered in this study, the thickest nitrided layer was obtained in solution D; thus, the sliding friction force test was carried out on an untreated sample and a sample from solution D to compare the abrasion resistance. The variation of the COF as a function of time for the untreated sample and that from solution D during non-lubricated wear tests are shown in Fig. 7. It shows that the COF of the sample from solution D, which was calculated from the average coefficient from the horizontal part of the curves, was reduced by approximately one third in the stable state compared with that of the untreated sample. In the initial stage of the curves, the COF of the sample from solution D reached the stable stage slower than the untreated sample. Fig. 8 shows the friction topography of the sample surface and 3D simulation diagram. It can be seen that the friction topography of the treated sample is more shallow and narrower than that of the untreated sample. The theoretical formula of the wear rate is based on the following equation:

K=$\frac{Weight_{wear}}{Distance_{total sliding}}$ (8)

Because the wear volume (obtained by multiplying the cross-section area of the sliding scratch by the sliding distance) times the density of the material is the wear loss, upon eliminating the same parameters, the result can be obtained by an analogue integration and comparison of the cross-section area. Upon completing the calculations, the cross-sectional area of the sample from solution D and the untreated samples were 2.25 × 10^-4 mm² and 6.70 × 10^-4 mm², respectively. The COF of the treated sample (0.49 ± 0.05) was nearly two-thirds that of the untreated sample (0.72 ± 0.03), and the wear loss rate of the untreated sample was nearly 3 times that of the treated sample. The results indicated that the wear resistance of the nitrided sample surface layer was obviously improved. Combining the results of Fig. 8, it can be inferred that the compounds of the nitrided layer played a major role in the wear resistance compared to the untreated layer. The nitrides and oxides in the compound layer improved the wear resistance of the samples, and the FeN_0.076 phase was able to reduce the COF [31]. Meanwhile, the distortion of the crystal lattice caused by grain boundary segregation of N atoms may enhance the wear resistance of the sample as a result of the water quenching after the nitriding process at a high voltage.

View original graphic| Download| PPT slide

Fig. 7.   Curves for COF vs. sliding time of the treated and untreated samples.

View original graphic| Download| PPT slide

Fig. 8.   Friction topography and 3D simulation diagram of (a) untreated sample and (b) the sample treated in solution D.

The cross-sectional Vickers hardness distribution of the treated samples is shown in Fig. 9. The hardness from the top surface towards the substrate was affected by the depth of the nitride layer. The hardness value of the nitride layer was approximately 400-450 HV_0.1, which is nearly 2.4 times the hardness of the substrate (180-210 HV_0.1). The thickness of the hardened region became broad with an increase in the urea concentration, which is in accordance with the distribution pattern of the thickness shown in Table 2. The steep gradients of the curves of Fig. 9 indicate the existence of an interface between the diffusion zone and substrate, which is consistent with the SEM morphologies in Fig. 5. Generally, the surface microhardness has an important connection with the phase composition of the layer. When considering the XRD patterns in Fig. 6, the relatively low hardness of the nitrided layers may be associated with the formation of a solid solution of nitrogen that increased the wear resistance of the material. Therefore, we assumed that when the working voltage was 190 V, the nitrogen atoms diffused into the austenite lattice to form a solid solution. Some saturated solid solutions precipitated nitrides as a result of the water quenching at the end of the nitriding process; however, when surrounded by a solid solution of nitrogen, the nitrided layer mainly existed in the form of γ_N [31].

View original graphic| Download| PPT slide

Fig. 9.   Hardness vs. depth curves of samples in treated in solutions of A, B, C and D.

3.4. Electrochemical measurements and analyses

Fig. 10 presents the potentiodynamic polarization curves of untreated and nitrided samples in 3.5% NaCl solution, and the obtained parameters are shown in Table 3, where i_corr and E_corr represent the corrosion current density and self-corrosion potential, respectively. In the case of untreated AISI 316L stainless steel, the polarization curve shows a passive region (-0.3 to 0.29 V_SCE) and a passive current density range (10^-6-10^-5 μA cm^-2). For all nitrided samples, the corrosion current densities were even lower compared to the untreated sample. For the nitrided samples treated in solutions A and B, the passivation regions were shorter than the untreated sample, and the corrosion current density was smaller in the sample from solution B. The polarization curve of the sample from solution A was similar to the substrate. The corrosion potential of the sample from solution B decreased approximately 500 mV compared to the substrate, but the corrosion current density decreased by an order of magnitude. The corrosion potentials of nitrided samples from solutions C and D increased by 100 mV and 300 mV compared to the substrate, respectively, and decreased passivation current densities on the order of 10^-7-10^-6 μA cm^-2 were measured. Therefore, upon an increase in the urea concentration of the nitriding solution, the corrosion resistance properties of nitrided layers were enhanced due to higher nitrogen content [[31], [32], [33], [34], [35], [36]].

View original graphic| Download| PPT slide

Fig. 10.   Potentiodynamic polarization curves of untreated and nitrided samples in 3.5% NaCl solution.

The corrosion resistance of nitrided layers may be related to the composition of the compound layer or the thickness of γ_N phase layer. From previous studies of AISI nitriding treatments on 316L steels [30,31], the mobilization of Cr leading to the formation of Cr_xN_y precipitates may cause a decrease in corrosion resistance at 470-480 °C. Considering the corrosion test results and the formation temperature of FeN_0.076 (500-520 °C), it can be confirmed that the precipitation of Cr_xN_y did not occur; thus, a negative effect for the corrosion resistance of nitrided layers on the AISI 316L steels was not observed.

4. Conclusions

In this work, CPEN with PEG2000 was applied to 316L stainless steel. The microstructure, phase composition, mechanical and corrosion resistance properties of the nitrided layer prepared in various urea solutions were investigated, and the main conclusions are as follows:

(1) From the current density-voltage (i_c-V) curves with PEG2000, it was found that i_spark decreased as U_spark improved, indicating that the addition of a macromolecular polymer provided sufficient energy for nitrogen atom infiltration due to the kinetics to realize rapid infiltration.

(2) The case depths obtained in different urea concentrations were different, and an increased urea concentration led to an increase in the thickness. With increasing urea concentration, the probability of surface oxidation increased, and the surface main component was Fe₃O₄.

(3) The results of the mechanical properties suggest that the material was greatly modified by the CPEN process, and the wear loss rate of the untreated sample was nearly 3 times higher than that of the treated sample. The COF of a sample in solution D was enhanced to ∼0.49, which was better than the untreated sample value of ∼0.79. The cross-section hardness values indicated that the main composition of the nitrided layer was the γ_N phrase.

(4) All the samples treated after the CPEN process had a high corrosion resistance in terms of the corrosion potential and corrosion current density. The improvement in corrosion resistance in solutions A and B was not obvious. Moreover, the samples in solutions C and D showed lower passivation current densities compared to the substrate.

Acknowledgements

This work was supported financially by the National Natural Science Foundation of China (No. 51771027), the Fundamental Research Funds for the Central Universities (No. FRF-BD-18-019A), the National Key Research and Development Program of China (No. 2017YFB0702100) and the National Environmental Corrosion Platform.